Power Platform Community Forum Thread Details

Hi all, connection issues are very disruptive for flows that use "Email with Options" and I'd like to hear any ideas to minimize this problem.

Our flow looks like a typical "approval" workflow using Email with Options.

- Triggered by users submitting a Microsoft Form.

- Updates a SharePoint List or Excel spreadsheet recording the form submission.

- Sends an email confirming we've received the form to the submitter.

- Sends an Email with Options to our approvals team which times out after 7 days.

- When approvers select an option, it's recorded to the List/Spreadsheet and a confirmation email is sent to the person who submitted the original form.

The problem is that after "fixing" any connection issue in Power Automate, the pending flows continue to use the old, invalid connection (recorded at the time the flow was triggered) instead of the newly authenticated connection.

For example, this morning we have 50 flows pending at the Email with Options step. We experienced a service disruption that caused sign in/authentication problems for users. I went to Power Automate and found, unsurprisingly, this also meant the "connections" our flow uses each needed to be fixed. I press the button on each to sign in again and those connections are up and running.

Now, new requests triggered from this moment will run fine. The 50 pending requests will receive responses over the next 7 days and then will fail because that run of the flow is using an old instance of the connection that's now invalid.

Other flows don't have this pending status halfway through. For other flows, you won't have as many failed runs (only those that were triggered during your outage instead of any triggered in the 7 days before it), and you can sensibly retry the failed runs. Retrying an Email with Options flow doesn't make business sense since EG you're resending the initial confirmation email, duplicating your approval email to approvers, and restarting your 7 day timer.

To process these runs in a timely manner, I have to spend the next 7 days re-checking for which flows have failed since my last check and make updates manually.

I've tried arranging a try-catch configuration for this flow so that we can handle errors thoughtfully. The problem is error handling also requires authorized connections (eg a connection to Outlook to send an email about the error to a flow owner) so the actions in my "catch" scope will fail to run.

Ideally any time a flow took a new action, it should use the flow's current connection authorization instead of whatever it recorded when the flow started. I think this would also be more secure. Has Microsoft commented on the likelihood of a change like this?

Instead let's discuss any strategies that could save us a headache when flows with long run times have connection issues. Does anyone have fixes or creative workarounds for this issue?

Categories:

Using flows